Big Data Baseball Page 20
While watching Cole in Indianapolis, Carroll rarely saw him throw curveballs, rather, fastballs, changeups, and occasional sliders. Carroll believes Fox has developed “a smarter version of a pitch count” and that the Pirates measure the workloads and stress levels of their pitchers using sophisticated mathematical models, favoring math over the eye test to decide when to rest pitchers or take them out of games.
Injury prevention is one of the final frontiers of big data. If a team could employ better injury-preventive practices, perhaps data-driven practices, and shave down even 10 percent on its injury rate of pitchers, it would gain an enormous competitive advantage.
Tampa Bay is believed to be one of the first teams to employ PITCHf/x data to monitor its pitchers’ health and predict and preempt injury. PITCHf/x tracks release points of pitchers, and a deviation from a normal release point may indicate injury. According to research from Jonah Keri’s book The Extra 2%, Tampa Bay had only one pitcher in the entire organization undergo Tommy John surgery from late 2005 to mid-2009.
The Pirates were believed to be gathering similar data in the summer of 2013 on Cole. They monitored what he threw and how much he threw. In his first 19 starts in the major leagues in 2013, Cole threw more than 100 pitches in a game only twice and threw fewer than 90 pitches in 6 of his starts. Several times in the second half of the 2013 season Cole either had his start pushed back several days or skipped, although the Pirates were not open about their specific inning, pitch, or workload limit for Cole.
“We do have some proprietary stuff we do in terms of workload. That information is made available to the coaches, and they have asked for various parts of it at various times. Not only starters but also relievers,” Fox said. “I don’t know how everything was arrived at [with the Cole situation], but I do know the stuff we provided was part of the process.”
So what was involved? Pitch types? High-stress innings? Total pitches?
“All of the above and some other [measurements],” Fox said. “A lot of the [preventive health models] work better with the more detailed information you have, so the PITCHf/x era is sort of the demarcation line.”
Clint Hurdle used some of that data with Cole and his other pitchers. Hurdle was careful with workloads. The Pirates ranked last in the major leagues in pitches per start in 2013, with their starting pitchers averaging just 90 pitches per appearance. Hurdle also tracked the bull-pen workload, consecutive days pitched, and how many times relievers were up and throwing in the bull pen.
While the Pirates lost several starters due to injury, their bull pen remained mostly healthy. The data approach reached beyond smarter pitch counts and analyzing PITCHf/x data such as velocity trends and release points. According to Carroll, in 2013 the Pirates were one of only twelve major league teams to put at least several of their prized arms through biomechanical evaluations, where pitchers were brought into a lab and had markers attached to their body. They went through their throwing motion and had the stress on their elbows and shoulders measured and compared to a database of healthy and unhealthy pitchers. But the biomechanical lab testing is imperfect as it is difficult for pitchers to throw normally while wearing a compression suit with dozens of electronic markers attached. Carroll, who writes for the Bleacher Report, reported in 2014 that the Pirates and the Baltimore Orioles were testing a compression sleeve from the company Motus. The sleeve fits easily over a pitcher’s throwing arm and elbow so he can throw naturally. The sleeve’s monitoring device is touted as being able to measure stresses on the elbow in real time and will perhaps be a breakthrough tool in injury prevention.
By 2014, the Pirates also had their minor league pitchers start to record their daily nutrition, hydration, rest, and workout routines. In spring training they began having their players wear Zephyr monitoring vests, on a voluntary basis. The tight-fitting compression shirt, which had a black, circular, detachable electronic device—about the size of a quarter—attached near the center of the chest, collected data from a sensor that records a player’s heartbeat and energy consumption.
Cole did not show any signs of fatigue as he kept pitching into September. He only got better. His performance data didn’t decline and his velocity was stable as his ability to prevent runs increased. His breakout performance was in win No. 82 against the Rangers, as he unleashed a sharp curveball that allowed him to strike out 9 batters.
Five days later against the Cubs, Cole was again masterful, allowing just one run while striking out seven in seven innings. Perhaps his most dominant outing came against the Padres on September 19, when he struck out a career-high 12 over six innings, allowing just four hits and a run. After that win the Pirates had moved to within a game of St. Louis. Cole won four straight September starts. Instead of being shut down at an arbitrary innings mark, he ended up being the club’s best pitcher in September, striking out 10.97 batters per 9 innings with a 1.69 ERA and a 4-0 record. The surge was tied to the increased use of his curveball. After throwing it sparingly early in his rookie campaign, he had nearly tripled his use of it to 20 percent of his pitches in September.
The Pirates never revealed what Cole’s red line was. Maybe he crossed it, or maybe he didn’t. Perhaps the human element—Hurdle’s observations—allowed Cole to cross that line. Or maybe the data-based workload monitors kept his arm fresh. Whatever the case, Cole kept pitching. He helped the Pirates continue to play great baseball into late September, propelling them to the cusp of a postseason berth.
* * *
September 23 was a cool night on the north side of Chicago. Temperatures dipped into the fifties. Autumn had arrived with a cool breeze coming off Lake Michigan. This season was rarely associated with important baseball in Pittsburgh. With a win over the Cubs the Pirates would clinch their first play-off appearance since 1992. A week earlier, Hurdle had sent out a daily e-mail message designed to mitigate any anxiety or mounting pressure:
“I never worry about the future. It comes soon enough.”
—Albert Einstein
Make a difference today.
Love Clint
Pirates starting pitcher Charlie Morton was matched up against the hard-throwing Jeff Samardzija, who had given the Pirates trouble in several meetings this season with a combination of his mid-90s fastball and hard-breaking curve. The matchup played out as expected, as a late-season pitchers’ duel.
Morton was excellent. Hitting his mid-to-low-90s sinker was akin to hitting a bowling ball, his opponents said. Batters thought the sinker seemed “heavier” than most other fastballs. The pitch induced ground ball after ground ball, and Morton threw seven shutout innings, allowing just three hits. He induced 11 groundouts against no fly balls.
Samardzija was nearly as good. He allowed one run over six innings as the Pirates took a 1–0 lead into the eighth inning. As the game neared an end, plastic sheeting akin to shower lining went up to protect lockers in the cramped visitors’ clubhouse at Wrigley Field, and twelve cases of champagne were wheeled in.
Hurdle removed Morton after he threw just 89 pitches. Setup man Mark Melancon would pitch the eighth with closer Jason Grilli ready for the ninth.
Melancon had been the game’s best setup man for the first five months of the season, but he had struggled in September. He had thrown a career-high number of cut fastballs—61.4 percent—and the pitch seemed to lack the same blistering effectiveness as it had earlier in the season. Brian Bogusevic singled off Melancon, a line drive to center field, to begin the bottom of the eighth. He advanced on a groundout and scored on a Donnie Murphy single to left field. Melancon got out of the inning without further damage, but the Cubs had tied the score at 1. For the third straight game the Pirates had allowed a lead to slip away after the seventh inning. Prior to these three games, the Pirates had won 76 of 77 games when leading after seven innings.
Kevin Gregg came into the game to pitch the top of the ninth for Chicago. With two outs and the bases empty, Starling Marte walked into the right-hander’s batter’s box.
He was back in the lineup but was still struggling from his hand injury in August. On a 2-1 pitch, Gregg threw a hanging slider that stayed up in the hitting zone. At that moment, Marte appeared to be completely healthy. He crushed the pitch. He knew it was gone as he took several steps toward first base, then punched both of his hands into the air. The ball landed in the sixth row in the left-field bleachers. Cheers rose from the several thousand dots of gold sitting in the cold, dark grandstand. Pirates fans had grown in number at road venues during the season’s second half. The Pirates took a 2–1 lead into the bottom of the ninth as Pirates closer Jason Grilli completed his warm-up throws in the makeshift bull pen that rests along the right-field line and then jogged toward the center of the diamond with a chance to pitch the Pirates into the postseason.
Grilli recorded two outs in the inning but also allowed a Nate Schierholtz single. Ryan Sweeney then came to the plate and rifled a Grilli fastball to the right-center-field gap. The Pirates’ outfield was playing deep to prevent an extra-base hit, and right fielder Marlon Byrd attempted to intercept the ball in deep right-center field, but failed to field it cleanly. As the ball deflected off Byrd’s glove and trickled away from him, Schierholtz sprinted around second base as Cubs third-base coach David Bell waved his arm violently like a propeller, urging Schierholtz to head toward home plate.
Pirates first baseman Justin Morneau saw Bell’s signal and tried to anticipate what would happen next. He drifted toward the middle of the infield. Morneau had not hit as the Pirates had hoped when they acquired him from the Twins at the end of August via a trade, but he remained one of the game’s better defensive first baseman. Morneau had been with the club less than a month, but at that moment he found himself the key man in a historic play.
“It’s probably not too often going to be a play at home on that kind of ball,” Morneau told reporters afterward. “I stay at first and try to keep [Sweeney there] and drift toward the middle once I see something different happen. Then I saw [the ball] kick away. That’s my cue there. I go back to the instincts. Something told me to go [to the center of the infield].”
In center field, Andrew McCutchen had moved over into a proper backup position and collected the ball that had deflected off Byrd’s glove. McCutchen was a great player but lacked one skill: a first-rate throwing arm. He had worked to strengthen it in the off-season through hundreds of long-toss sessions with his Florida neighbor and former Pirates player Steve Pearce. McCutchen fielded the ball and made an accurate one-bounce throw to Morneau near the pitcher’s mound. Morneau collected the toss and quickly pivoted and delivered a perfect throw to Pirates catcher Russell Martin, who was blocking the plate.
Martin had played the second half of the season with a bad left knee, and he was the last line of resistance between clinching a play-off berth or going to extra innings. Martin took the throw from Morneau and hung on to the ball as Schierholtz collided with him. On his knees, minus his catcher’s mask, which he had tossed away for a better view of the action, Martin lifted the ball triumphantly above his head. Home-plate umpire Mike DiMuro punched the air to signal the final out. The moment was captured by Associated Press photographer Charles Rex Arbogast and became an iconic image of the season. The game was over. The Pirates were in the play-offs.
When the Texas Rangers had clinched a postseason berth several years earlier, star outfielder Josh Hamilton, who was battling alcoholism, asked the team to celebrate with alcohol-free beverages for fear a champagne-and-beer celebration could trigger a relapse. As the likelihood of the Pirates clinching a play-off berth had approached, several veteran players had approached Hurdle and asked him if he preferred that the team celebrate with nonalcoholic beverages. His battle with alcoholism was not a secret, and he had not had a sip of alcohol in fourteen years. The visitors’ clubhouse at Wrigley Field was the smallest in the game, and twelve cases of champagne would not only drench the worn carpet but fill the cramped confines with a mist of alcohol.
“I told them it didn’t bother me,” Hurdle told the Pittsburgh Post-Gazette earlier in the day. “I wasn’t going to lick my lips. I’m way past that. I’m a grown man.… I know what I can’t do. I could be a drunk tomorrow if I had one beer. Like I always say, one is too many and a million aren’t enough. I wanted to embrace the moment. I wanted to be a part of all of it. I wanted to get wet. I wanted to be soaked. I wanted it dripping down my face. I wanted my eyes to sting.”
For twenty minutes after the final out, Hurdle and the Pirates participated in a baseball ritual that hadn’t happened in a Pittsburgh-occupied clubhouse since the Pirates captured the National League’s East Division in 1992. They doused themselves with champagne and cheap beer. They wore eye protection of every variety, from ski goggles to dollar-store swimming goggles to the bulky eye protection you’ll commonly find in an introduction-to-chemistry classroom. The motley crew of veteran cast-offs, bonus-baby draft picks, old-school coaches, and mathematical wizards had got there by becoming a sum greater than that of their individual parts. They had to accept new voices, new ideas, and collaboration. That was their story and their accomplishment in 2013. And by the end of September, they weren’t ready for the ride to end.
12
MAGIC ACT
Black-and-white math blended with the various colors of subjective observations and insights several hours before each series. In 2013, prior to the first home game of every series, Dan Fox and Mike Fitzgerald sat in on the strategy meeting Clint Hurdle held in his office at PNC Park, along with the usual crowd of assistant coaches and video coordinators. In every such meeting that season, Fox and Fitzgerald were present. Beginning in 2013, in addition to preseries meetings before home games, Hurdle included Fox or Fitzgerald on a conference call during pre-series meetings on the road, and Fitzgerald even began traveling with the team on a number of road trips. While nearly every club had hired at least one nerdy mathematical genius number cruncher, by 2013 few if any such quantitative analysts had been integrated into meetings and even the team’s traveling party as Fox and Fitzgerald were with the Pirates.
The Pirates had let their quantitative analysts out from where some might have imagined them to reside, a basement surrounded by computer servers. They were allowed to enter the ego-filled, exclusive atmosphere of a major league clubhouse. And what was most remarkable was Fox and Fitzgerald were accepted there. In 2013, a communication barrier, and even a lack of respect, still often existed between old-school and new-school camps throughout the sport. Quantitative analysts and on-field staff and players were often very different people, with very different backgrounds, temperaments, and prejudices. One common complaint heard from analytics staffers in the game is that their data-based findings do not always reach the field. But this complaint was not heard in Pittsburgh.
Since the middle of the 2012 season, the Pirates’ analytics team had begun to play more and more of a role. Fox’s and Fitzgerald’s job responsibilities grew beyond making off-season evaluations on potential player acquisitions. They were useful in more than just implementing big-picture strategies, such as shifting, identifying the value of pitch framing, or finding inefficiencies in the draft. Their roles extended beyond macro-level analysis. In 2013, the Pirates leaned on their analysts more than ever for help at the micro level, the daily game-planning.
“It started with Clint really reaching out and saying, ‘Hey, I want you guys to be more involved and be here more in terms of the meetings,’” Fox said.
Hours before the start of every series is a general meeting with Hurdle and his assistants, as is the case with most major league teams. Then specialized meetings follow: the hitting coach goes over scouting reports and video with his hitters; the pitching coach does the same with his starters and relievers; and the infield and outfield coaching staffs meet with the position players. But before those specialized meetings, preparation against an opponent starts with a comprehensive scouting and strategy-setting meeting in Hurdle’s office, where members of the major leag
ue coaching staff, advance video scout Wyatt Toregas, and someone like special assistant to the general manager Jim Benedict, who is on the road scouting the next opponent, meet. In 2012, Fox or Fitzgerald were integrated into opponent-scouting meetings at the start of all home series, which often comprise multiple series. Come 2013, they were each involved in every preseries meeting, home and road.
“[Before 2012] it was more e-mail-type exchanges. [In 2013] we were talking every three days, opposed to once during the homestand,” Fox said. “That leads to just hanging around, more face time, more off-the-cuff conversations. In every little conversation is something we can learn from and help with. [Trust] I think is tied to time and exposure.”
In part because of Hurdle’s openness and increasing trust bred through familiarity, Fox and Fitzgerald counseled Hurdle on day-to-day lineup construction and bull-pen usage in addition to defensive alignment. Hurdle also encouraged his assistant staff and players to ask questions of Fox and Fitzgerald. Some players, such as reliever Mark Melancon, frequently interacted with Fitzgerald in the clubhouse. “There were some skeptics early on and even now,” Fitzgerald said of the clubhouse’s view of PITCHf/x data. “So it’s good to be [available].” Fitzgerald didn’t preach sabermetrics ideology to players, he did not try to convert nonbelievers, nor did he try to explain WAR—unless asked. Rather, Fitzgerald made himself available. Over time in a clubhouse wary of outsiders, as most major league clubhouses are, he became a familiar face. In these meetings, particularly the preseries meetings with Hurdle, the art of subjective, observational scouting—such as getting a feel for a player’s recent mechanical changes or weaknesses, or even mental temperament—met science, the streams of data flowing from the Pirates’ proprietary database.